课程: MLOps with Databricks

免费学习该课程！

今天就开通帐号，24,600 门业界名师课程任您挑！

Deploy serving endpoints

Deploy serving endpoints - Databricks教程

课程: MLOps with Databricks

开始一个月试用 购买团队方案

Deploy serving endpoints

“

- [Instructor] In this video, I will be talking about deployment of the serving endpoints. Until now, we successfully registered three models in unit two catalog and created the richest pack. Let's go ahead and deploy three serving endpoints, hotel-cancellations-model for the basic model, hotel-cancellations-model-fe for the model with feature lookup, hotel-cancellation-preds for the feature spec called return predictions. All the serving endpoints can be deployed in the same fashion no matter what gets served behind it. Using the Databricks as decay workspace client. Let's initialize an instance of the workspace client class called workspace. We use workspace serving endpoints create methods to create the hotel-cancellations-model endpoint. It requires to input a name and the config for the endpoint. As part of the configuration we need to provide served entities. It's a list that consists of served entity input class instances. This list can contain up to 15 entities, and traffic…

内容